Hypothesis Testing

Introduction to Hypothesis Testing

Hypothesis testing is a fundamental procedure in statistics used to evaluate claims about population parameters, such as proportions, means, or variances. This chapter introduces the key concepts, terminology, and step-by-step procedures for conducting formal hypothesis tests.

• Hypothesis: A claim or statement about a property of a population.

• Hypothesis Test (Test of Significance): A procedure for testing a claim about a property of a population using sample data.

Key Components of Hypothesis Testing

Hypothesis tests involve several general steps that apply to tests about proportions, means, or standard deviations/variances.

• Null Hypothesis (H0): States that the value of a population parameter is equal to a claimed value.

• Alternative Hypothesis (H1 or Ha): States that the parameter differs from the null hypothesis, using symbols <, >, or ≠.

• Significance Level (α): The probability threshold for determining when sample evidence is significant enough to reject H0. Common choices: 0.05, 0.01, 0.10.

• Test Statistic: A value calculated from sample data, used to decide whether to reject H0. Examples: z, t, χ2.

• P-value: The probability of obtaining a test statistic at least as extreme as the observed value, assuming H0 is true.

• Critical Region: The set of values for the test statistic that leads to rejection of H0.

Types of Hypothesis Tests

• Two-tailed test: Critical region in both tails of the distribution; used when Ha uses ≠.

• Left-tailed test: Critical region in the left tail; used when Ha uses <.

• Right-tailed test: Critical region in the right tail; used when Ha uses >.

Step-by-Step Procedure for Hypothesis Testing

The following steps outline the formal process for conducting a hypothesis test:

1. Identify the claim and express it in symbolic form.

2. State the symbolic form that must be true if the original claim is false.

3. Formulate hypotheses:

   • Alternative hypothesis H1: Symbolic form without equality (<, >, ≠).

   • Null hypothesis H0: Symbolic form with equality (=).

4. Select the significance level (α): Commonly 0.05.

5. Identify the test statistic and its sampling distribution (normal, t, χ2), based on the parameter and sample size.

6. Calculate the test statistic and determine the P-value or critical value(s).

7. Make a decision:

   • If P-value < α, reject H0.

   • If P-value > α, fail to reject H0.

   • Alternatively, compare the test statistic to the critical value(s).

8. Restate the decision in simple, nontechnical terms, addressing the original claim.

Example: Testing a Claim About a Proportion

Suppose we want to test the claim that "most Internet users utilize two-factor authentication," i.e., the population proportion p > 0.5. With a sample of n = 926 and observed sample proportion &hat;p = 0.52:

• Step 1: Claim: p > 0.5

• Step 2: If claim is false: p ≤ 0.5

• Step 3: H1: p > 0.5; H0: p = 0.5

• Step 4: Select α = 0.05

• Step 5: Use normal distribution (z-test) since np ≥ 5 and nq ≥ 5

• Step 6: Calculate test statistic:

• Step 7: Find P-value (right-tailed): P-value ≈ 0.106

• Step 8: Since P-value > α, fail to reject H0. There is not sufficient evidence to support the claim.

Test Statistics and Sampling Distributions

The choice of test statistic and sampling distribution depends on the parameter being tested and sample size.

P-Value and Critical Value Methods

• P-Value Method: Compare P-value to α.

• Critical Value Method: Compare test statistic to critical value(s) that define the rejection region.

For right-tailed tests, the critical region is to the right of the critical value; for left-tailed, to the left; for two-tailed, in both extremes.

Type I and Type II Errors

Errors can occur in hypothesis testing:

• Type I Error: Rejecting H0 when it is true. Probability = α.

• Type II Error: Failing to reject H0 when it is false. Probability = β.

Power of a Hypothesis Test

The power of a test is the probability of correctly rejecting a false null hypothesis, calculated as 1 - β. Higher power means a greater chance of detecting a true effect.

• Power increases as the difference between the true parameter and the null hypothesis value increases.

• Commonly, a power of at least 0.80 is desired in experimental design.

Confidence Intervals and Hypothesis Testing

A confidence interval provides a range of likely values for a population parameter. If a claimed value is not within the confidence interval, it can be rejected. For means and variances, confidence intervals and hypothesis tests usually lead to the same conclusion; for proportions, they may differ.

Restating Decisions in Nontechnical Terms

Final conclusions should be stated simply, addressing the original claim. Avoid saying "accept the null hypothesis"; instead, say "fail to reject the null hypothesis." This indicates insufficient evidence to reject H0, not proof that H0 is true.

• Example: "There is not sufficient evidence to support the claim that most Internet users utilize two-factor authentication."

Summary

• Hypothesis testing is a structured process for evaluating claims about population parameters.

• Key steps include formulating hypotheses, selecting significance level, calculating test statistics, and making decisions based on P-values or critical values.

• Understanding errors, power, and confidence intervals is essential for interpreting results.